Romosozumab, a humanized monoclonal antibody that binds and inhibits sclerostin, produces a marked increase in bone formation with a concomitant decreased bone resorption. This transient rise in bone formation in the first 2 months of treatment is mainly due to an increased modeling-based bone formation. This requires the recruitment and differentiation of osteoblasts, one possibility being a preferential switch in commitment of precursors to osteoblasts over adipocytes. The purpose of this study was to analyze the marrow adiposity in transiliac bone biopsies at months 2 or 12 from the FRAME biopsy sub-study in patients receiving romosozumab or placebo. The total adipocyte area, number, and density were measured on the total cancellous bone area. The size and shape at the individual adipocyte level were assessed including the mean adipocyte area, perimeter, min and max diameters, and aspect ratio. No significant difference in total adipocyte area, number, or density between placebo and romosozumab groups was observed at months 2 and 12, and no difference was observed between 2 and 12 months. After 2 or 12 months, romosozumab did not modify the size or shape of the adipocytes. No relationship between the adipocyte parameters and the dynamic parameters of bone formation could be evidenced. In conclusion, based on the analysis of a small number of biopsies, no effect of romosozumab on bone marrow adiposity of iliac crest was identified after 2 and 12 months suggesting that the modeling-based formation observed at month 2 was not due to a preferential commitment of the precursor to osteoblast over adipocyte cell lines but may result from a reactivation of bone lining cells and from a progenitor pool independent of the marrow adipocyte population.